Light generated or reflected from objects existing in nature may have an associated inherent value in wavelength or range of wavelengths or similar measurement. An image sensor may be a device that takes the images of objects. Some image sensors may implement semiconductor devices that react with external energy (e.g. light) to generate reproducible electrical signals. Particularly, pixels of the image sensor may sense the light generated from objects and convert the sensed light into electric signals or values.
Two types of semiconductor image sensors are CCDs (Charge Coupled Devices) based on a silicon semiconductor substrate and a CMOS image sensor manufactured by sub-micro CMOS (Complementary Metal Oxide Semiconductor) processing technology.
A CCD is a device in which MOS capacitors may be positioned very close to each other and charge carriers may be stored and carried to capacitors. CCDs may be driven in a relatively complicated manner and may consume a relatively high amount of power, among other potential limitations. Formation of a CCD may require a relatively large number of steps using a mask processes. Accordingly, it may be relatively difficult to implement a signal processing circuit in a CCD chip as compared to a CMOS image sensor (e.g. the ability to remove or minimize defects).
A CMOS image sensor may have a PD (Photo Diode) and a MOS transistor in unit cells and may implement an image by detecting signals in a switching method. A CMOS image sensor may have a relatively low manufacturing cost and power consumption compared to a CCD and may be relatively easily integrated with a peripheral chip. As described above, a CMOS image sensor may be manufactured by CMOS technology, such that it may be easily integrated with a peripheral system for amplifying and processing a signal. Therefore, it may be possible to minimize the manufacturing cost by implementation of CMOS image sensors. Further, the processing speed of a CMOS image sensor may be relatively high and the power consumption may be relatively low compared to a CCD image sensor (e.g. about 1% of the power consumption of a CCD in some applications).
An image sensor may be formed by ion-injecting photo diodes into a semiconductor substrate. The size of the photo diodes may be minimized to maximize the number of pixels in a unit area without increasing the chip size and the overall area of a photodetection unit may be minimized. However, a stack height (e.g. thickness) may not easily be minimized, which may present challenges in producing and designing a backside-illuminated image sensor (e.g. ability to minimize the step at the upper portion of the illumination unit and/or minimize interference with light due to metal routing).
In the backside-illuminated image sensor of the related art, when electrons generated by irradiated light are collected to corresponding pixels, accurate light reception characteristics may be exhibited. However, since there is no layer or structure which is able to isolate electrons generated by light irradiated on the back side between pixels, electrons may be unintentionally transferred between adjacent pixels, which may result in undesirable crosstalk.